# Normalization

> Source: https://sukruyusufkaya.com/en/glossary/normalizasyon
> Updated: 2026-05-13T20:58:53.949Z
> Type: glossary
> Category: veri-bilimi-ve-veri-yonetimi
**TLDR:** The process of bringing numerical variables to a defined scale to make them more suitable for modeling and comparison.

<p>Normalization aims to place numerical variables with different ranges onto a common scale, making model learning more balanced. Distance-based algorithms, neural networks, and optimization-driven models are often sensitive to scale differences. It is undesirable for a feature with large magnitude to dominate others purely because of its scale. For this reason, techniques such as min-max normalization are widely used. However, the decision to normalize should always consider the data distribution and the type of model being used.</p>